Tom Anioł

✨ open source project without marketing, is just a private repository with public access ✨

Yes

1 engine. 1,000+ stores. 0 platform fees per tenant. That's the promise of multi-tenant commerce & the reality if you do it right. This guide covers how modern multi-tenancy systems work and how to implement them on composable engines. Download here → rigbyjs.com/resources/mult…

We're starting an awesome project with @kuchniavikinga

We took everything we've learned from building eCommerce and marketplaces, and we rebuilt our entire delivery process from scratch. AI-Native. We poured that knowledge of building 600+ custom commerce features into our own framework for delivering software. 👇

Everyone's testing AI design tools on landing pages and simple dashboards. I wanted to test something harder: can AI design complex B2B e-commerce? Not just make it look good. Actually understand what should be on the page, plan the information architecture, and handle the complexity of B2B flows. One prompt. Four tools. Same test. Prompt: "Build a B2B e-commerce store for industrial MRO supplies (maintenance, repair, operations). Plan what the store needs and design it." Intentionally simple. No feature list. No wireframes. No hints about what pages to include or what elements matter. I wanted to see how much each tool understands about e-commerce on its own. Not just design skill, but commerce logic. The tools: Google Stitch, v0 (Vercel), Figma Make, Claude. Results: 🏆 v0 @v0 (Vercel) Clear winner. v0 was the only tool that stopped before designing. It asked clarifying questions, proposed a plan with pages and feature scope, and waited for confirmation. Then it delivered the best layout: proper category navigation, bulk pricing, trust signals, quote request flows. And under the hood: a clean Next.js project with reusable components. Real code structure you can actually continue building on. 🔵 Google Stitch @stitchbygoogle Most original visual style. Stitch generates designs that don't look like typical AI output. It has its own aesthetic, makes interesting layout decisions, and feels fresh. The problem: click "View Code" and you get a single flat HTML file with Tailwind classes. No components. No structure. No modularity. You'd have to decompose everything into proper components before using any of it in production. For ideation and moodboarding it's great. For building, not yet. One thing I really like though: Google introduced DESIGN.md with Stitch. It's a markdown file that captures your entire design system in natural language plus specific values (colors, spacing, typography, component styles). You can export it and hand it off to Claude Code, Cursor, or any coding agent. I've been building something similar in my own workflow, but never had one standardized format for it. Smart move from Google and worth watching closely. 🟣 Figma Make @figma Similar layout approach to v0, but noticeably weaker execution. Feels like v0 from a year ago. The upside: from most of these tools you can export or transfer designs into Figma for further refinement. ⬛ Claude @claudeai Technical, dark, dense. No visual hierarchy. As a standalone design-from-scratch tool, not useful in this form. But this is not how I use Claude. And this is where it gets interesting. Our production flow at @Rigbyjs is Claude Code paired with our own e-commerce component library. We've spent the last two years building and battle-testing these components across 30+ @medusajs stores. Every component is proven in production, maintained, and continuously improved. When Claude Code works on top of that library with our rules and conventions, the output quality is on a completely different level. It understands the constraints, follows the patterns, and delivers working e-commerce that we can ship with confidence. No other tool comes close to this today. I think this is the real unlock for building e-commerce with AI: not generating designs from zero, but having a strong, tested component foundation that AI can build on top of. The better your library, the better the AI output. That compounds over time. Bottom line: for going from zero to design with real B2B commerce understanding, v0 wins. Best planning, best design, best code output. For production builds, nothing beats Claude Code with a purpose-built component library. That's where the real speed and quality lives.

A generation ago, Poland rationed sugar and flour while its citizens were paid one-tenth what West Germans earned. Today its economy has edged past Switzerland to become the world’s 20th largest with over $1 trillion in annual output. nbcnews.com/business/econo…

@thekitze in da house showing @openclaw use cases on @AITinkerers Poland 🔥

Come join us and welcome @thekitze as special guest of next AI Tinkerers Poland meetup 🙌

@thekitze 🥶

my living costs of living in poland €1200 - vitamin d supply €1400 - antidepressants €1200 - therapy €990 - electricity for running 10k lux lights everywhere to simulate sunlight) €3900 - ai credits €2400 - printer filament help me budget this my family is starving

I mean.. If coding costs go to zero.. the real effort is to stand out from the crowd 🤔

OK I have to admit I was surprised to meet engineers and EMs today who work in the Warsaw office of: - Google - NVIDIA - Netflix - Snowflake - Meta - Superhuman - Wix Can’t name another city in Europe save for London with this much concentration of Big Tech + hot startups! 🇵🇱

@BrendanFoody @felicis @benchmark @generalcatalyst Congrats! Your chart gives me hope for @Mercurjs PS: pls don't sue me

We’ve raised our $350M Series C at a $10B valuation from @felicis, @benchmark, and @generalcatalyst. Just 2 years after starting, Mercor is paying $1.5 million per day to experts in our marketplace. We’re creating a new category of work in the AI economy, where software engineers, bankers, lawyers, and other professionals earn based on their experience while advancing the frontier of AI. While most new categories take time to build momentum, we’ve broken every growth record. For comparison, in their first 2 years: - Uber paid out just over a $1 million to drivers - Airbnb paid out $10 million to hosts We are unlocking human potential in the AI economy.

One advantage of having many distinct sovereign nations instead of a single world government is that they can try different policies and we can see which work best.

Damn... I could have thought about @ProductHunt campaign today for @Rigbyjs 😑 Nonetheless we are focused building hardest builds in for #eCommerce and #marketplaces No time for watching #RegularShow :(

@moustache_doman @elonmusk Jakbyś miał jednego odsprzedać... ✋

Say I'm in Poland and I want to purchase 500 starlinks mini. Whom should I talk to? @elonmusk

🇵🇱 Poland is nearly as rich as the UK.

@adam__xyz True, I think this trend works for a16z & Founder's Fund (fund next big breakthroughs), also for foundational LLMs companies like Claude video (accelerate next big thing), but Kalshi is just trading/gambling on the outcomes of real-world events. It's a centralised Polymarket

getting tired of software promo videos that compare their work to athletic excellence (carl lewis), space exploration, political breakthroughs (always the berlin wall lol) and rock icons (freddie mercury). besides it being a tired trend (a16z's first was amazing), you are doing a disservice to freddie, you didn't write "bohemian rhapsody", you hunch over slack all day selling secondaries under the delusion that building software is somehow rebellious, i'd love to see you at wembley saying one original thing

@thekitze @grabbou 🔥🔥🔥

THE NATION OF POLAND IS NOW CARRYING REACT (AND NATIVE) ON ITS BACK POLSKA STRONK 🇵🇱

One of the most interesting projects of 2025 for @Rigbyjs 🚀